Xref: news2.mv.net comp.os.msdos.djgpp:2737
From: Eli Zaretskii <eliz AT is DOT elta DOT co DOT il>
Newsgroups: comp.os.msdos.djgpp
Subject: Re: End of file keyboard symbol.
Date: Sun, 14 Apr 1996 20:29:31 +0200
Organization: NetVision LTD.
Lines: 24
Message-ID: <Pine.SUN.3.91.960414202247.3098H-100000@is>
References: <316E0DA2 DOT 38EF AT dwnplaza DOT ncom DOT nt DOT gov DOT au>
NNTP-Posting-Host: is.elta.co.il
Mime-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
In-Reply-To: <316E0DA2.38EF@dwnplaza.ncom.nt.gov.au> 
To: djgpp AT delorie DOT com
DJ-Gateway: from newsgroup comp.os.msdos.djgpp


On Fri, 12 Apr 1996, Dale Robinson wrote:

>   while(scanf("%d",&n) != EOF) {
>     if (n < 0) {
>       printf("warning: n is -ve\n");
>     } else {
>       printf("%d is correct\n", n);
>     }

Never assume that `scanf' returns EOF when it hits an end of file.  You 
should in general test for the return value to be the expected number of 
items you want to read (in this case 1), and if the returned value is 
less than that, take an error action.

Btw, the current version of `scanf' has a bug that causes it to NEVER 
return EOF, but even after that bug is corrected, relying on EOF is 
unsafe.  For instance if you try to read 2 items, like this:

	 scanf ("%d %d", &i, &j);

and press ^Z after typing only one number, ANSI C says that `scanf' 
shouldn't return EOF, but 1.  In general, it only has to return EOF when 
it hits end of file ``before any conversion takes place''.